PromptRAID
PromptRAID is a free, open-source, self-hosted command-line framework for red-teaming prompt injection in LLM agents.
What PromptRAID is
PromptRAID is a free, open-source, self-hosted command-line framework for red-teaming prompt injection in LLM agents. Its mutation engine creates payload variants using encoding changes, jailbreak or role-play framing, delimiter breaks, and indirect instructions delivered through tool or document output. Implemented mutation categories map to verified MITRE ATLAS technique IDs, while three additional categories are unverified placeholders. Triggered prompt injection and data leakage appear in the taxonomy but currently have no mutation generators. The harness supports Anthropic, Gemini, Groq, Cerebras, and OpenRouter, including OpenAI-compatible endpoints. A judge labels outcomes as success, partial compliance, fail, or inconclusive, distinguishing neutral quoting from adopting an injected instruction. Runs and per-payload results are stored in SQLite, with JSON and HTML report exports. The CLI includes taxonomy, mutation, judging, test, demo, and report commands. PromptRAID requires Python 3.9 or later and at least one provider API key. Shikhali Jamalzade is identified as its author.
Good
- Maps implemented tests to MITRE ATLAS
- Generates varied injection payloads
- Supports five provider options
- JSON and HTML report exports
- MIT licensed and self-hosted
Watch out
- Data leakage has no mutation generator
- Triggered injection has no generator
- Three taxonomy categories are unverified
- Requires Python 3.9 or later
Who it suits
It suits offensive security researchers and teams performing prompt-injection red-team exercises on LLM agents. Users should be comfortable with self-hosted command-line tools and provider credentials.
Verdict
PromptRAID provides a focused CLI workflow for generating, running, judging, and reporting prompt-injection tests. Its taxonomy mapping and provider coverage are useful for security work, while several listed categories remain incomplete.
The facts
Each one read from PromptRAID's own pages, with the date it was checked.
What it does
- ATLAS coverage
- Implemented mutation categories map to verified MITRE ATLAS technique IDs, while some listed categories remain unverified placeholders or have no generator yet.checked 9 Oct 2026
- Installation
- The README recommends installing the CLI with pipx and also documents installation in a Python virtual environment.checked 10 Oct 2026
- Intended users
- The project describes its intended use as systematic red-team testing and benchmarking of LLM-based agents across providers.checked 10 Oct 2026
- Interface
- The project provides a command-line interface with commands for taxonomy, mutation, judging, running tests, demos, and reports.checked 9 Oct 2026
- License
- PromptRAID is distributed under the MIT License, which permits use, copying, modification, distribution, sublicensing, and sale subject to its terms.checked 10 Oct 2026
- Limitations
- The README says triggered prompt injection and data leakage appear in its taxonomy but do not yet have mutation generators.checked 10 Oct 2026
- Maintainer
- The README identifies Shikhali Jamalzade (alisalive) as an independent offensive security researcher and the project author.checked 10 Oct 2026
- Model providers
- The harness supports Anthropic, Gemini, Groq, Cerebras, and OpenRouter, including OpenAI-compatible endpoints.checked 9 Oct 2026
- Payloads
- Its mutation engine generates prompt injection variants using encoding tricks, role-play framing, delimiter breaking, and indirect injection through tool or document output.checked 10 Oct 2026
- Payload testing
- Its mutation engine generates prompt injection variants using encoding tricks, jailbreak framing, delimiter breaking, and indirect injection through tool or document output.checked 9 Oct 2026
- Provider credentials
- Running tests requires at least one provider API key, with environment variables documented for Anthropic, Gemini, Groq, OpenRouter, and Cerebras.checked 10 Oct 2026
- Providers
- The harness supports Anthropic, Gemini, Groq, Cerebras, and OpenRouter, including OpenAI-compatible endpoints.checked 10 Oct 2026
- Purpose
- PromptRAID is an automated prompt injection red-teaming framework for LLM-based agents, mapped to MITRE ATLAS.checked 10 Oct 2026
- Reports and storage
- PromptRAID exports JSON and HTML reports and stores runs and per-payload results in SQLite.checked 10 Oct 2026
- Requirements
- The package requires Python 3.9 or later and asks users to configure at least one supported provider API key.checked 9 Oct 2026
- Security classification
- The project maps prompt injection tests to MITRE ATLAS and describes them as red-team tests for LLM agents.checked 10 Oct 2026
- Taxonomy
- Mutation categories map to MITRE ATLAS technique IDs, which the project says it checks against the live ATLAS matrix.checked 10 Oct 2026
- Unverified categories
- Three additional categories are described as unverified placeholders and are excluded from the listed ATLAS mapping.checked 10 Oct 2026
- Verdicts
- The judge classifies results as success, partial compliance, fail, or inconclusive, and distinguishes neutral quoting from adopting an injected instruction.checked 10 Oct 2026
- Verdict scoring
- The judge classifies results as success, partial compliance, fail, or inconclusive and distinguishes neutral echo from adopted compliance.checked 9 Oct 2026
- Warranty
- The MIT License provides the software “as is” and disclaims warranties and liability.checked 9 Oct 2026
The company
- Maker
- The repository names Shikhali Jamalzade (alisalive) as the author.checked 9 Oct 2026
Alternatives to PromptRAID
- 1
OpenSecureAI Scanner
Free planOpen sourceWebAPILinuxmacOS8.1From $49 a month - 2
PromptGuard
Free planWebAPIExtensionLinux7.9From $19 a month - 3
Promptfoo
Free planOpen sourceWebAPILinuxmacOS7.8Free plan - 4
Giskard
Free planOpen sourceWebAPILinuxSelf-hosted7.6Free plan - 5
AIRTA Red Team
Free planOpen sourceWebAPILinuxmacOS7.5Free plan - 6
CyberSecEval
Free planOpen sourceWebAPILinuxmacOS7.5Free plan